Validity of biopsy-based drug effects in a diet-induced obese mouse model of biopsy-confirmed NASH

Compounds in clinical development for nonalcoholic steatohepatitis (NASH) improve liver histopathology in diet-induced obese mouse models of biopsy-confirmed NASH. Since the biopsy section used for histopathological evaluation represents only < 1% of the whole mouse liver, we evaluated how well biopsy-based quantitative image analyses correlate to stereology-based whole-liver quantitative changes upon drug treatment.

Male leptin-deficient Lep^ob/Lep^ob mice were fed the Amylin liver NASH (AMLN) diet for 16 weeks before stratification into treatment groups using a biopsy-based evaluation of type I collagen αI (col1a1) levels. Mice were treated for 8 weeks with either vehicle (PO, QD), liraglutide (0.4 mg/kg, SC, QD), elafibranor (30 mg/kg, PO, QD) or INT-767 (10 mg/kg, PO, QD). Terminal quantitative histological assessment of liver lipid (hematoxylin-eosin staining), inflammation (galectin-3 immunohistochemistry (IHC); gal-3), and fibrosis (col1a1 IHC) was performed on terminal liver biopsies and compared with stereologically sampled serial sections spanning the medial, left and right lateral lobe of the liver.

The distribution of liver lipid and fibrosis was markedly consistent across lobes, whereas inflammation showed some variability. While INT-767 and liraglutide significantly reduced total liver weight by 20 and 48%, respectively, elafibranor tended to exacerbate hepatomegaly in Lep^ob/Lep^ob-NASH mice. All three compounds markedly reduced biopsy-based relative liver lipid content. Elafibranor and INT-767 significantly reduced biopsy-based relative gal-3 levels (P < 0.001), whereas INT-767 and liraglutide tended to reduce relative col1a1 levels. When changes in liver weight was accounted for, both INT-767 and liraglutide significantly reduced biopsy-based total col1a1 content. Although minor differences in absolute and relative liver lipid, inflammation and fibrosis levels were observed across lobes, the interpretation of drug-induced effects were consistent with biopsy-based conclusions. Notably, the incorporation of changes in total liver mass revealed that liraglutide's efficacy reached statistical significances for all analyzed parameters.

In conclusion, in-depth analyses of liver homogeneity demonstrated that drug-induced improvement in liver biopsy-assessed histopathology is representative for overall liver effects assessed using stereology. Importantly, these findings reveal how changes in whole-liver mass should be considered to provide a deeper understanding of apparent drug treatment efficacy in preclinical NASH studies.